AlphaFold reveals the structure of the protein universe (deepmind.com)

Article summary

DeepMind's AlphaFold AI system has been used to predict the 3D structure of nearly all catalogued proteins known to science, expanding the AlphaFold Protein Structure Database to over 200 million structures. This update has the potential to dramatically increase our understanding of biology and accelerate progress on important real-world problems. The AlphaFold DB is a searchable database that is open and free to all, and has already been accessed by over half a million researchers. The predicted structures can be used to advance research on various issues, including sustainability, food insecurity, and neglected diseases.

Main themes

• protein structure prediction
• AlphaFold AI system
• biological complexity
• machine learning in biology
• protein dynamics and interactions
• scientific research and discovery

What commenters say

• The accuracy of AlphaFold's predictions is uncertain and can only be verified through experimental methods like x-ray crystallography.
• AlphaFold's predictions are based on existing known structures, which may not be representative of all proteins, particularly non-globular ones.
• The release of AlphaFold's predictions is a significant advancement in the field of protein structure prediction, but it is not a solution to the 'protein universe' problem.
• The complexity of biological systems, including protein dynamics and interactions, is still not fully understood and requires further research.
• AlphaFold's predictions can be used as a starting point for further research and experimentation, but should not be taken as ground truth.
• The diversity and complexity of biological systems, including proteins, make them more challenging to understand than human-engineered systems.
• Other research teams have also made significant progress in protein structure prediction using machine learning techniques, and the field is becoming increasingly competitive.